Evidence of the cloud lifetime effect from wildfire-induced thunderstorms

Transcription

1 GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L22809, doi: /2008gl035680, 2008 Evidence of the cloud lifetime effect from wildfire-induced thunderstorms Daniel T. Lindsey 1,2 and Michael Fromm 3 Received 12 August 2008; revised 7 October 2008; accepted 14 October 2008; published 25 November [1] A case study is presented of pyro-cumulonimbi (pyrocbs) forming over Canadian forest fires. Cloud-top ice effective radius values of these pyrocbs are significantly smaller than are those within surrounding convection. The smoke provides a massive source of cloud condensation nuclei (CCN), resulting in smaller cloud droplets which freeze homogeneously at temperatures around 40 C and produce very small ice crystals. It is also shown that the pyrocb anvils persist 6 12 hours longer than convectivelygenerated cirrus anvils from nearby convection. This provides evidence for the so-called cloud lifetime effect, an aerosol indirect effect identified by the most recent Intergovernmental Panel on Climate Change (IPCC) report. Citation: Lindsey, D. T., and M. Fromm (2008), Evidence of the cloud lifetime effect from wildfire-induced thunderstorms, Geophys. Res. Lett., 35, L22809, doi: /2008gl Introduction [2] The Intergovernmental Panel on Climate Change Fourth Assessment Report [Intergovernmental Panel on Climate Change, 2007] identified the cloud lifetime effect, also referred to as the aerosol second indirect effect, as a mechanism by which aerosols modify the microphysical and thereby the resulting radiative properties of clouds. The majority of literature addressing the cloud lifetime effect has focused on low stratiform cloudiness, while relatively less attention has been given to cirrus clouds. Lohmann and Feichter [2005] noted that only small effects on cirrus clouds were observed following the eruption of Mt. Pinatubo. Conversely, Koren et al. [2005] observed differences in convective cloud properties between areas with relatively large and small aerosol optical depths over the Atlantic Ocean. Dirtier conditions were associated with convective anvils having smaller ice effective radius, an increased cloud fraction, and a larger area. Given that these results were based on a statistical study, the authors note that it remains uncertain whether the aerosols were directly affecting cloud properties. Sherwood [2002] concludes that tropical convective anvils have smaller ice crystals in regions with more aerosols. He suggests that the aerosols serve as cloud condensation nuclei (CCN), making smaller cloud droplets which then freeze homogeneously, resulting in smaller ice crystals. However, like Koren et al. [2005], 1 Regional and Mesoscale Meteorology Branch, National Environmental Satellite, Data, and Information Service, NOAA, Fort Collins, Colorado, USA. 2 Cooperative Institute for Research in the Atmosphere, Colorado State University, Fort Collins, Colorado, USA. 3 Naval Research Laboratory, Washington, D. C., USA. This paper is not subject to U.S. copyright. Published in 2008 by the American Geophysical Union. Sherwood [2005] establishes a statistical relationship between aerosols and anvil ice size and shows no direct evidence of the responsible physical mechanism. [3] Massie et al. [2007] used Moderate Resolution Imaging Spectroradiometer (MODIS) data to look for aerosol indirect effects of clouds at different heights. They found that cloud reflectance increased with increasing aerosol optical depth (AOD) for low-level clouds, but did not increase for larger AOD values for high-level ice clouds. Their motivation for examining cloud reflectance was partially due to Twomey s [1977] finding that more aerosols will result in more reflective clouds (the first indirect effect). Massie et al. [2007] conclude...for the MODIS observations in February and March that aerosol indirect effects are not present for pressure levels associated with ice crystals... This statement is valid for the first indirect effect, but does not necessarily hold for the cloud lifetime effect. In addition, they looked at all ice clouds and made no attempt to discriminate between convectively generated anvils. [4] Large forest fires provide an ideal natural experiment to study the effect of aerosols on convective cloud microphysics. Storms forming directly over fires are commonly referred to as pyro-cumulonimbi, or pyrocbs, [e.g., Fromm et al., 2005]; Rosenfeld et al. [2007] observed significantly smaller ice effective radii between a pyrocb and surrounding convection. Fromm et al. [2008] showed an individual pyrocb anvil that persisted for 1.5 days. In this paper, we present an example of several Canadian pyrocbs which not only produce significantly smaller ice crystals than nearby convection, but also generate anvils which persist longer than those anvils having larger ice. 2. GOES Observations of Canadian PyroCbs [5] The Geostationary Observational Environmental Satellite s (GOES) band 2 (3.9 mm) is ideal for locating large fires, even at relatively high latitudes [Weaver et al., 2004]. In fact, the steep viewing angle often allows for fires to be located which might not be visible at lower latitudes since we can sometimes see beneath a cloud forming directly over a fire. Figure 1 shows the GOES mm image from 0100 UTC on 5 July 1998 over British Columbia and the Yukon Territory. The color scale is designed such that hot spots (i.e., fires) show up as white pixels. A number of wildfires are evident in northwestern British Columbia and throughout southern Yukon Territory. In the forthcoming 10.7 mm satellite imagery, all pixels whose 3.9 mm brightness temperature exceeded 37 C at 0100 UTC are denoted by black dots so that the wildfire locations can be easily compared to the clouds. L of5

2 Figure 1. GOES mm image from 5 July 1998 at 0100 UTC over northern British Columbia and the Yukon Territory. The color bar is in C, so the white dots represent hot spots. The blue lines are latitude and longitude in degrees north and degrees west, respectively. [6] Figure 2 shows a 6-panel display of GOES mm imagery from 4 5 July 1998 over the same domain as in Figure 1, and a full loop from 2100 UTC on 4 July to 2000 UTC on 5 July is available as Animation S1 in the auxiliary material. 1 At 2100 UTC (Figure 2a) convection has formed over the higher terrain, but most of the area to the west of the mountains near the wildfires remains cloudfree. By 0200 UTC (Figure 2b) a few storms have begun to form over fires; the most obvious of these pyrocbs are circled. The most isolated pyrocb anvil is quite evident in northern British Columbia at 0830 UTC (Figure 2c), while a larger, more diffuse group of pyrocb anvil clouds can be seen in central Yukon Territory. Note also that most of the regular convection which formed over the higher terrain has dissipated by this time. At 1530 UTC (Figure 2d) these pyrocb anvils persist, but by 1845 UTC (Figure 2e) the southernmost anvil has begun to thin as indicated by its brightness temperatures warming above 30 C (the cutoff between white and yellow in the color scale; anvil remnants denoted). Conversely, the northern pyrocb anvil persists and still has brightness temperatures colder than 30 C at 2200 UTC on 5 July (Figure 2f). [7] By making use of the solar reflected component of the GOES 3.9 mm band, cloud-top ice effective radius estimates can be obtained. Using the retrieval of Lindsey and Grasso [2008], Figure 3 shows the ice effective radius at 0200 UTC on 5 July, while Animation S2 provides a loop from 1800 UTC on 4 July to 0200 UTC on 5 July. The pyrocb anvils that were noted in Figure 2b have effective radius values at the low end of the retrieval s capability, near 7 mm, while the surrounding convection has values ranging from about 15 to 40 mm. In monitoring this effective radius retrieval over the last two summers over North America, the authors have rarely witnessed values as low as 7 mm except in the case of pyrocbs. In this 4 July 1998 case, we can 1 Auxiliary materials are available in the HTML. doi: / 2008GL safely conclude that cloud-top ice sizes produced by pyrocbs are significantly smaller than by other nearby convection. [8] The most interesting aspect of this event is related to the relative longevity of the pyrocb anvils compared to the regular convective anvils. In Figure 2, note how the two pyrocb anvils still have brightness temperatures colder than 52 C (the change from magenta to turquoise in the color bar represents 52 C) at 0830 UTC, while the surrounding convection has mostly either dissipated or thinned enough to allow for warmer brightness temperatures. By 1530 UTC (Figure 2d) the pyrocb-generated anvils are the only ones remaining with brightness temperatures colder than 30 C. The southern pyrocb finally dissipates by 1830 UTC, but the northern pyrocb-generated anvil still remains at 2200 UTC, nearly 24 hours after it originally formed. [9] Table 1 lists the approximate formation and dissipation times of the anvils generated by pyrocbs and representative surrounding convection. Formation and dissipation are defined as when the 10.7 mm brightness temperature first cools/warms below/above 30 C, respectively. The Southern PyroCb refers to the southernmost anvil circled in Figure 2b, and the Northern PyroCbs refer to the group of anvils which form in the Yukon Territory and whose anvils coalesce by 0830 UTC (Figure 2b). In defining the Regular Convection, the initiation and dissipation times are estimated because the anvils from many storms forming at slightly different times quickly become a contiguous entity. The Southern Regular storms refer to those forming in north-central British Columbia prior to 2100 UTC on 4 July (Figure 2a), while the Northern Regular storms are those which form in east-central Yukon Territory (Figure 2a). As indicated in Table 1, anvils from the regular convection last around 12 hours, while those from the pyrocbs persist from 18 hours (southern) to 29 hours (northern). 3. Discussion [10] Without detailed in situ measurements of wildfire aerosols and the resulting pyrocbs, it is impossible to conclude definitively that the particles affect the storms microphysics. In the case of these 4 5 July 1998 storms, however, the evidence obtained from GOES measurements is rather convincing that the smoke aerosols directly affected the convection. Biomass fires generate a significant number of particles which are efficient CCN, while ice nuclei (IN) generation is relatively less active (P. Demott, manuscript in preparation, 2008). By adding massive numbers of CCN, cloud droplets will be significantly smaller since the available water vapor will be spread out among a larger number of particles. These cloud droplets are then lofted by the updrafts of convective storms to temperatures colder than 40 C and frozen homogeneously, resulting in very small ice crystals [Rosenfeld et al., 2007]. Figure 3 provides convincing evidence that pyrocbs produce smaller cloudtop ice crystals than the surrounding convection. [11] Andreae et al. [2004] found that smoke from wildfires leads to smaller convective cloud droplets, reducing precipitation efficiency and allowing for more cloud mass to enter a convective storm s anvil. A larger ice mass in the anvil of the pyrocbs observed on 4 5 July 1998 is 2of5

3 Figure 2. GOES mm images from 1998 over British Columbia and the Yukon Territory at (a) 2100 UTC on 4 July, (b) 0200 UTC on 5 July, (c) 0830 UTC on 5 July, (d) 1530 UTC on 5 July, (e) 1845 UTC on 5 July, and (f) 2200 UTC on 5 July over the same domain as in Figure 1. Black dots represent pixels whose 3.9 mm brightness temperature at 0100 UTC on 5 July 1998 exceeded 37 C. Locations of PyroCb anvils are denoted. presumably the explanation for their significantly longer cloud lifetimes. With more ice, the process of sublimation and settling simply takes longer in the pyrocb-generated anvils. A possible conclusion from these observations is that more CCN leads to smaller anvil ice crystal size, which prolongs the anvil s lifetime. [12] The pyrocb anvils on 5 July contained smoke. Indeed the absorbing Aerosol Index (AI) from NASA s Total Ozone Mapping Spectrometer (TOMS) on that day exceeded any other in Canada in July [Fromm et al., 2000, Figure 5]. Having now identified strong pyrocbs on 4 July we contend that this was the causal event for the stratospheric smoke revealed by Fromm et al. [2000, Figure 2]. Figure 4 provides GOES-9 visible imagery from 5 July Figure 3. GOES-9 Ice Effective Radius from 0200 UTC on 5 July 1998 over the same domain as in Figure 1. Colors indicate effective radius (mm) for all pixels whose 10.7 mm brightness temperature is colder than 40 C. Warmer pixels are assigned a black to white color according to their 10.7 mm brightness temperature. Table 1. Anvil Lifetime of PyroCbs and Regular Convection on 4 6 July 1998 Anvil Formation Time Dissipation Time Duration Southern PyroCbs 0030 UTC 4 July 1830 UTC 5 July 18 hr Northern PyroCbs 2316 UTC 4 July 0600 UTC 6 July hr Southern Regular 1915 UTC 4 July 0715 UTC 5 July 12 hr Northern Regular 1915 UTC 4 July 0830 UTC 5 July 13 hr 3of5

4 Figure 4. GOES-9 visible images from (a) 1300 UTC, (b) 1600 UTC, (c) 1830, and (d) 2330 UTC on 5 July 1998, over the same domain as in Figure 1. In Figure 4c, each red X denotes those pixels whose Aerosol Index (AI), from NASA s Total Ozone Mapping Spectrometer (TOMS), exceed , and Animation S3 shows a visible loop from 2100 UTC on 4 July to 2000 UTC on 5 July. Although somewhat difficult to observe from four snapshots, Animation S3 shows that the optically thick anvil from the southern pyrocb persists until after sunrise on 5 July. Insolation is impeded under this anvil, delaying convective cloudiness until later in the day compared to surrounding areas. This observation shows the effect wildfire smoke had on the surface radiation over a fairly large area. [13] Early morning view/zenith angles accentuate the anvils and remnants in Figures 4a and 4b, the northern one arcing over northern Yukon into Northwest Territories, the southern one (straddling the British Columbia/Yukon border) with smooth texture and sharp shadowing. In Figure 4c (late morning), each red X denotes those pixels whose AI exceeded 3 at 1830 UTC on 5 July. Values greater than 3 signify areas with great absorbing aerosol amount and/or altitude [Fromm et al., 2008]. These pixels correspond to the anvil remnants from both the northern and southern 4 July pyrocbs. Thus it is strongly evident that these day-after high-altitude anvils contain a mixture of evaporating ice crystals and forest fire smoke from multiple pyrocbs. [14] An increase in low clouds results in a net negative top of the atmosphere (TOA) radiative forcing, while the rise in reflected solar radiation associated with an increase in high clouds may be offset by longwave radiative cooling [Albrecht, 1989]. However, any increase in cloudiness, regardless of cloud level, will decrease solar radiation at the surface. Additionally, optically thick convective anvils are generally surrounded by a ring of optically thinner clouds [Koren et al., 2005] which have a more robust net TOA radiative effect. The overall radiative effect of an increase in thick cirrus clouds is a very complex issue (particularly when absorbing aerosols are contained within the cloud), and is beyond the scope of this paper. From a single case study, we cannot conclude that an increase in aerosols always results in an increased lifetime of convective anvils. However, this example provides convincing evidence that the cloud lifetime effect is valid not only for low stratiform clouds, but also for convective ice anvils, and demands that more research be conducted. A logical next step would be to seek an inverse correlation between cloudtop ice effective radius and cloud lifetime over a season or multiple years. It is possible that a massive influx of CCN, such as from a wildfire, is needed to extend the lifetime of convective clouds. [15] Acknowledgments. This material is based on work supported by the National Oceanic and Atmospheric Administration under grant NA17RJ1228. The authors would like to thank Mark DeMaria and John Knaff for valuable comments, as well as two anonymous reviewers. The views, opinions, and findings in this report are those of the authors, and should not be construed as an official NOAA and or U.S. Government position, policy, or decision. References Albrecht, B. A. (1989), Aerosols, cloud microphysics, and fractional cloudiness, Science, 245, Andreae, M. O., D. Rosenfeld, P. Artaxo, A. A. Costa, G. P. Frank, K. M. Longo, and M. A. F. Silvas-Dias (2004), Smoking rain clouds over the Amazon, Science, 303, Fromm, M., J. Alfred, K. Hoppel, J. Hornstein, R. Bevilacqua, E. Shettle, R. Servranckx, Z. Li, and B. Stocks (2000), Observations of boreal forest fire smoke in the stratosphere by POAM III, SAGE II, and lidar in 1998, Geophys. Res. Lett., 27, Fromm, M., R. Bevilacqua, R. Servranckx, J. Rosen, J. P. Thayer, J. Herman, and D. Larko (2005), Pyro-cumulonimbus injection of smoke to the stratosphere: Observations and impact of a super blowup in northwestern Canada on 3 4 August 1998, J. Geophys. Res., 110, D08205, doi: /2004jd Fromm, M., O. Torres, D. Diner, D. Lindsey, B. Vant Hull, R. Servranckx, E. P. Shettle, and Z. Li (2008), Stratospheric impact of the Chisholm pyrocumulonimbus eruption: 1. Earth-viewing satellite perspective, J. Geophys. Res., 113, D08202, doi: /2007jd Intergovernmental Panel on Climate Change (2007), Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, edited by S. Solomon et al., Cambridge Univ. Press, Cambridge, U. K. 4of5

5 Koren, I., Y. J. Kaufman, D. Rosenfeld, L. A. Remer, and Y. Rudich (2005), Aerosol invigoration and restructuring of Atlantic convective clouds, Geophys. Res. Lett., 32, L14828, doi: /2005gl Lindsey, D. T., and L. Grasso (2008), An effective radius retrieval for thick ice clouds using GOES, J. Appl. Meteorol. Climatol., 47, Lohmann, U., and J. Feichter (2005), Global indirect aerosol effects: A review, Atmos. Chem. Phys., 5, Massie, S. T., A. Heymsfield, C. Schmitt, D. Muller, and P. Seifert (2007), Aerosol indirect effects as a function of cloud top pressure, J. Geophys. Res., 112, D06202, doi: /2006jd Rosenfeld, D., M. Fromm, J. Trentmann, G. Luderer, M. Andreae, and R. Servranckx (2007), The Chisholm firestorm: Observed microstructure, precipitation, and lightning activity of a pyro-cumulonimbus, Atmos. Chem. Phys., 7, Sherwood, S. C. (2002), Aerosols and ice particle size in tropical cumulonimbus, J. Clim., 15, Twomey, S. (1977), The influence of pollution on the shortwave albedo of clouds, J. Atmos. Sci., 34, Weaver, J. F., D. T. Lindsey, D. E. Bikos, C. C. Schmidt, and E. Prins (2004), Fire detection using GOES-11 rapid scan imagery, Weather Forecast., 19, M. Fromm, Naval Research Laboratory, Code 7227, 4555 Overlook Avenue SW, Washington, DC 20375, USA. D. T. Lindsey, Cooperative Institute for Research in the Atmosphere, Colorado State University, 1375 Campus Delivery, Fort Collins, CO , USA. (lindsey@cira.colostate.edu) 5of5